The verdict: WordPress.com’s MCP support is the most significant thing to happen to automated publishing since the REST API — and most people are sleeping on it.
Automattic quietly shipped an MCP server that lets AI agents like Claude and ChatGPT connect directly to your WordPress.com site. Create posts, edit pages, manage content, publish — all from a conversation with an AI. No custom scripts. No API wrangling. Just configure the MCP connection and your AI agent has full publishing power.
This is either the future of content or the beginning of an SEO spam apocalypse. Probably both. Let’s break it down.
What Actually Happened
Automattic — the company behind WordPress.com — built an official MCP (Model Context Protocol) server for WordPress sites. MCP is the open standard Anthropic created that lets AI models talk to external tools through a standardised interface. Think of it as USB-C for AI integrations.
With MCP support, any compatible AI agent can now:
- Create and publish blog posts directly to your site
- Edit existing content — update old posts, fix metadata, rewrite sections
- Manage site structure — categories, tags, pages, media
- Read site data — pull analytics, check what’s published, audit content
The setup is straightforward. In Claude Desktop or Claude Code, you add the WordPress MCP server to your configuration, authenticate with your WordPress.com credentials, and you’re connected. ChatGPT works similarly after OpenAI adopted MCP support in early 2025.
No custom plugins. No REST API tokens to manage manually. No MU-plugins to maintain. It just works.
Why Developers Should Care
If you’ve ever built a publishing pipeline — and we literally built one for this blog — you know the pain. Custom API integrations, authentication token management, Yoast SEO field mapping, block format conversion. Hours of plumbing work before you can publish a single post programmatically.
MCP collapses that entire stack into a configuration file.
That’s not a small deal. For indie devs running content sites, for agencies managing client blogs, for anyone who’s automated their publishing workflow with custom scripts — this is a genuine simplification of the toolchain.
The practical win: you can now go from “research a topic” to “published on WordPress” in a single AI conversation. No intermediate scripts. No deployment pipeline. Just intent to published post.
The Content Creator Angle
For non-technical content creators, this is even bigger. You no longer need a developer to set up automated publishing. Connect Claude to your WordPress site, describe what you want, and it writes and publishes.
The workflow looks like this:
- Tell Claude your topic and target keywords
- Claude researches, outlines, and drafts the post
- You review and approve (or don’t — your call)
- Claude publishes directly to WordPress with proper formatting, categories, and metadata
That’s a one-person content operation that would’ve required a writer, an editor, and a webmaster five years ago. Now it’s a conversation.
But here’s where we need to be honest about what this actually enables.
The Elephant in the Room: AI Spam at Scale
WordPress powers roughly 43% of the web. Giving every WordPress.com site owner a one-click AI publishing pipeline is going to produce an ocean of mediocre content. That’s not speculation — it’s arithmetic.
The barrier to publishing AI-generated content just dropped to zero. Before MCP, you needed technical skills to automate publishing. Now you need a WordPress.com account and a Claude subscription. That’s it.
Google says AI content is fine if it’s helpful. But “helpful” is doing a lot of heavy lifting in that sentence. Most AI-generated content isn’t helpful — it’s competent filler that reads like every other AI-generated post on the same topic.
The winners will be the people who use AI agents as a force multiplier for genuine expertise — not a replacement for having something to say.
If you’re using Claude to publish content about topics you actually understand, with real opinions, real data, and real value — MCP makes your workflow dramatically faster. If you’re using it to spray 50 generic posts a week across your WordPress sites, Google will eventually notice. They always do.
MCP vs REST API: Which Should You Use?
If you already have a custom WordPress REST API integration — like we do — the question is whether to migrate.
Use MCP if:
- You’re starting fresh and want the fastest path to automated publishing
- You’re on WordPress.com (not self-hosted WordPress.org)
- You want your AI agent to handle the full workflow end-to-end
- You don’t need deep plugin-specific integrations (Yoast SEO fields, custom post types)
Stick with REST API if:
- You’re self-hosted on WordPress.org
- You need granular control over Yoast SEO metadata, custom fields, or plugin-specific features
- You’ve already built a working pipeline and it’s doing the job
- You need operations MCP doesn’t expose yet
The honest answer: MCP is the future, but the REST API is more mature for power users. We’re keeping our custom integration for now because Yoast SEO field control matters for what we do. But for most people starting today, MCP is the right call.
How to Set It Up (5 Minutes)
The setup is almost insultingly simple:
- Get your WordPress.com site — if you don’t have one, the free tier works
- Install Claude Desktop (or use Claude Code)
- Add the WordPress MCP server to your Claude MCP configuration — point it at your site, authenticate with OAuth
- Start a conversation — “Write a blog post about [topic] and publish it to my site”
That’s it. No npm packages. No Python scripts. No environment variables to configure. Claude handles the rest.
For ChatGPT users, the process is similar — OpenAI’s MCP integration lets you connect the same WordPress MCP server and publish through their interface.
What This Means for the Next 12 Months
Here’s our prediction: MCP-powered publishing becomes the default for new WordPress content sites within a year. The convenience gap is too large. Building custom API integrations will feel like writing raw SQL when an ORM exists.
The second-order effect is more interesting. When publishing is frictionless, the bottleneck shifts entirely to content quality and strategy. The technical moat around automated publishing disappears. Everyone gets the same tools. The differentiator becomes what you publish, not how you publish it.
That’s actually good news for people who have something real to say. And terrible news for content farms that relied on the technical barrier to entry keeping competition manageable.
The verdict stands: WordPress MCP support is a genuine inflection point. Use it to publish faster, but don’t mistake speed for strategy. The AI can publish — but it still can’t think for you.
Building an AI-powered content pipeline? We’ve been doing it for months. The tools are the easy part — the hard part is having opinions worth publishing.